Prompts used by the Ghostbuster paper, taken from the official repo. 2k prompts for creative writing and long context generation. Mainly used by the paper to benchmark LLM detection, but could be useful for benchmarking many other things (coherence, factuality, creative writing, etc.).
